Infrastructure, modelled for the duration

Cash flows are contractual but not certain. Availability payments, regulated returns and concession revenues each have their own risk profile. Inflation indexation, change-in-law clauses and counterparty creditworthiness all affect realised yield.

ESG is now LP table stakes. Carbon intensity, biodiversity impact and social licence reporting are increasingly demanded by institutional LPs. Manual collection does not scale across a portfolio.

Hold periods outlast deal teams. When the partner who underwrote the asset has moved on, the platform is the institutional memory. Decision provenance is non-negotiable.

Refinancing is the second underwriting. Most infrastructure assets refinance multiple times during the hold. Each refinancing is a fresh underwriting against a new rate environment.

Infrastructure carries 14 sub-asset overlays in the Reuben AI rubric registry: Toll roads, Airports, Seaports, Rail and rolling stock, Bus and transit, Water and wastewater utilities, Electricity transmission and distribution, Gas networks, District heating, Telecom towers, Fibre networks, Satellite ground infrastructure, Social infrastructure PPP and Waste to energy. Those overlays are part of the 314 sub-asset overlays the platform maintains across 56 of its 68 native asset classes.

An overlay never reweights the parent rubric. It adds the extra scoring criteria, risk flags, document types and key term fields that only apply to that slice, so a toll roads deal and a waste to energy deal are both scored on the shared infrastructure rubric while each is still asked the questions specific to its structure. That is what keeps deals inside one asset class comparable to each other and to the rest of the portfolio.

Selecting an overlay at intake changes what the platform expects: the document checklist, the extraction schema for key terms, and the flags raised when a term sits outside the range the rest of your infrastructure book has accepted. Nothing is retrofitted later by hand.

Every infrastructure opportunity that moves through Reuben AI leaves a structured record rather than a folder of documents and a memo. The rubric score is stored with the criteria that produced it, each extracted key term is stored with the clause and page it came from, and every flag is stored with the reason it fired and who cleared it.

The practical effect is that the investment committee paper, the LP report line and the audit trail are all reading the same underlying record. When an LP asks in year three why a infrastructure position was underwritten the way it was, the answer is retrieved, not reconstructed.

Infrastructure sits as a core allocation for Infrastructure, Family Office and Institutional mandates, and is adjacent or opportunistic for 3 of the other fund types the platform serves.

That matters operationally because most funds do not hold infrastructure alone. The moment a portfolio spans more than one asset class, comparability becomes the constraint: a position scored on a bespoke infrastructure spreadsheet cannot be ranked against a position scored somewhere else. Running infrastructure on the same rubric engine as the rest of the book removes that break.

Reporting inherits the same property. Portfolio monitoring, valuation cadence, capital account movement and LP reporting all draw from the single stored record, so a infrastructure position appears in the quarterly pack on the same basis as every other holding, with the same evidence behind it.